Dehazing using Color-Lines
RAANAN FATTAL
The Hebrew University of Jerusalem
Photographs of hazy scenes typically suffer from having low-contrast and
offer a limited visibility of the scene. This paper describes a new method for
single-image dehazing that relies on a generic regularity in natural images
where pixels of small image patches typically exhibit a one-dimensional
distribution in RGB color space, known as color-lines. We derive a local for-
mation model that explains the color-lines in the context of hazy scenes and
use it for recovering the scene transmission based on the lines’ offset from
the origin. The lack of a dominant color-line inside a patch or its lack of
consistency with the formation model allows us to identify and avoid false
predictions. Thus, unlike existing approaches that follow their assumptions
across the entire image, our algorithm validates its hypotheses and obtains
more reliable estimates where possible.
In addition, we describe a Markov random field model which is dedicated
for producing complete and regularized transmission maps given noisy and
scattered estimates. U nlike traditional field models that consist of local cou-
pling, the new model is augmented with long-range connections between
pixels of similar attributes. T hese connections allow our algorithm to prop-
erly resolve the transmission in isolated regions where nearby pixels do not
offer relevant information.
An extensive evaluation of our method over different types of images
and its comparison to state-of-the-art methods over established benchmark
images shows a consistent improvement in the accuracy of the estimated
scene transmission and recovered haze-free radiances.
Categories and Subject Descriptors: I.3.3 [Computer Graphics]: Pic-
ture/Image Generation—Display algorithms; I.4.1 [Computer Graphics]:
Digitization and Image Capture—Radiometry
Additional Key Words and Phrases: image dehazing, contrast enhancement,
transmission estimation
ACM Reference Format:
Raanan Fattal 20013. Dehazing using Color Lines. ACM Trans. Graph. 28,
4, A rticle 106 (August 2009), 11 pages.
DOI = 10.1145/1559755.1559763
http://doi.acm.org/10.1145/1559755.1559763
Raanan Fattal acknowledges the Israeli Science Foundation (ISF) for s up-
porting this research.
Author’s addresses: raananf@cs.huji.ac.il.
Permission to make digital or hard copies of part or all of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
show this notice on the first page or initial screen of a display along with
the full citation. Copyrights for components of this work owned by others
than ACM must be honored. Abstracting with credit is permitted. To copy
otherwise, to republish, to post on servers, to redistribute to lists, or to use
any component of this work in other works requires prior specific permis-
sion and/or a fee. Permissions may be requested from Publications Dept.,
ACM, Inc., 2 Penn Plaza, Suite 701, New York, NY 10121-0701 USA, fax
+1 (212) 869-0481, or permissions@acm.org.
c
2009 ACM 0730-0301/2009/14-ART106 $10.00
DOI 10.1145/1559755.1559763
http://doi.acm.org/10.1145/1559755.1559763
1. INTRODUCTION
Small dust particl es or liquid droplets in the air, collectively known
as aerosols, scatter the light in the atmosphere. This light deflection
reduces the direct scene transmission and replaces it with a layer
of previously-scattered ambient light known as airlight or veiling
light. Consequently, photographs taken in hazy or dusty weather
conditions, and even ones taken in relatively clear days but capture
long distances, are often of low-contrast and offer a limited visibil-
ity of the scene. A similar difficulty is encountered in underwater
photography.
Most image dehazing methods remove the layer of haze by re-
covering the direct scene radiance. These methods rely on a physi-
cal image formation model that describes the hazy image as a con-
vex combination between the scene radiance and the atmospheric
light. As we detail in Section 3, the coefficients of this linear com-
bination correspond to the scene transmission (visibility) at each
image pixel. In case of RGB images, this model consists of four
unknowns per pixel, the scene radiance at each color channel and
the transmission value, whereas the input image supplies only three
constraints, the intensity of each channel.
In order to resolve this indeterminacy many methods require ad-
ditional information about the scene, such as multiple images taken
at different weather conditions [Narasimhan and Nayar 2000] or
polarization angles [Schechner et al . 2001] and knowing the scene
geometry [Kopf et al. 2008]. More recently, methods that alleviate
these input requirements were developed. This is achieved either by
relaxing the physical model, for example by seeking for an image of
maximal contrast [Tan 2008], or by introducing additional assump-
tions over hazy scenes. For example, Fattal [2008] resolves the in-
determinacy by assuming a local lack of correlation between the
transmission and surface shading functions. While this approach
is capable of providing physically-consistent estimates, it cannot
be applied at regions where the two functions do not vary suffi-
ciently. He et al. [2009] robustly estimate the transmission from
pixels with a dark (low-intensity) color channel. This approach re-
quires that such pixels are found across the entire image. Large re-
gions of bright surfaces in the image bias towards under-estimated
transmission.
In this paper we propose a new method for single-image de-
hazing that takes advantage of a generic regularity in natural im-
ages in w hich pixels of small image patches typically exhibit one-
dimensional distributions in RGB color space, known as color-
lines [Omer and Werman 2004]. We use this observation to define a
local image formation model that reasons the color-lines in the con-
text of hazy images and allows recovering the scene transmission
based on the lines’ off set from the origin. Moreover, the unique
pixel distribution predicted by the formation model allows us to
identify patches that do not exhibit proper color-lines and discard
them. In contrast to existing approaches that follow their assump-
tions across the entire image, our algorithm validates its hypothe-
ses and hence obtains more reliable transmission estimates where
possible. In this paper we focus on esti mating the transmission ac-
curately and assume the at mospheric light vector is given.
ACM Transactions on Graphics, Vol. 28, No. 4, Articl e 106, Publicati on date: August 2009.
